Search Results

The TROJAN-ENZ® brand condom is used for contraception and for prophylactic purposes (to help prevent pregnancy and the transmission of sexually transmitted diseases).

The condoms are made of a natural rubber latex sheath, which completely covers the penis with a closely fitted membrane. The condoms are smooth surface straight-walled nipple-end (SWNE) style within ASTM standard specifications D-3492 Table 1 requirements, e.g., minimum length 160 mm, maximum width 54 mm, and minimum thickness of 30 μM.

This 510(k) notification is for a male latex condom, a Class II medical device, and therefore does not include a study proving device effectiveness in the same way an AI/ML medical device might. The focus of this submission is on demonstrating substantial equivalence to existing predicate devices, primarily by meeting established industry standards for condom manufacturing.

Here's an analysis of the provided information in the context of your request for AI/ML device studies:

1. Table of Acceptance Criteria and Reported Device Performance

Explanation: For this type of device, meeting industry standards (like ASTM D-3492) and demonstrating material and design equivalence to previously approved devices are the primary "acceptance criteria" and "performance" metrics. The ASTM standard specifies various physical properties necessary for the device to perform its intended function (e.g., burst strength, freedom from holes, dimensions).

Regarding the AI/ML-specific questions (2-9), this document does not contain any information because it's for a traditional medical device (latex condom) and not an AI/ML-driven device.

Here's why each is not applicable and what would typically be expected for an AI/ML device:

2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)

• Not Applicable: This is a physical device, not an AI/ML algorithm. There's no "test set" of data in the AI/ML sense.
• For AI/ML: A test set would comprise a distinct, unseen dataset used to evaluate the final model's performance. Details like size, demographics, clinical setting, and whether data was collected specifically for the study (prospective) or from existing records (retrospective) are crucial.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g. radiologist with 10 years of experience)

• Not Applicable: No ground truth establishment by experts is described for a latex condom's physical properties.
• For AI/ML: Ground truth for diagnostic AI often requires multiple expert reviewers to label data. Their number, specialty, and experience heavily influence the credibility of the ground truth.

4. Adjudication method (e.g. 2+1, 3+1, none) for the test set

• Not Applicable: No test set or expert adjudication is relevant here.
• For AI/ML: Adjudication describes how disagreements among experts are resolved (e.g., 2 experts agree, 1 provides tie-breaker; 3 experts, majority rules; or independent consensus).

5. If a multi reader multi case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance

• Not Applicable: This is not an AI-assisted diagnostic or classification device.
• For AI/ML: An MRMC study is common for AI tools designed to assist human readers (e.g., radiologists). It measures the change in human performance (e.g., sensitivity, specificity, AUC) when aided by the AI compared to human performance alone, often quantifying the "effect size" of the improvement.

6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done

• Not Applicable: Not an AI algorithm.
• For AI/ML: Standalone performance evaluates the AI algorithm's ability to perform a task independently, without human intervention. This is relevant for autonomous AI systems.

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc)

• Not Applicable: Ground truth for a physical product typically relies on physical measurements and material testing against established standards.
• For AI/ML: The "gold standard" or ground truth can vary. For medical images, it might be histopathology results, long-term follow-up/outcomes, or the consensus of multiple highly experienced specialists.

8. The sample size for the training set

• Not Applicable: No AI model, thus no training set.
• For AI/ML: The training set is the data used to teach the AI model. Its size is a critical factor in model performance, with larger and more diverse sets generally leading to better generalization.

9. How the ground truth for the training set was established

• Not Applicable: No AI model or training set.
• For AI/ML: Similar to the ground truth for the test set, the method of establishing labels for the training data (e.g., single expert, multiple experts, automated labeling, linkage to pathology) is essential for understanding the model's foundation.

In summary, this 510(k) notification for a male latex condom demonstrates compliance through adherence to recognized international standards (ASTM D-3492) and substantial equivalence to legally marketed predicate devices, rather than through AI/ML-specific performance studies.

Ask a specific question about this device